Part of the problem is that inter-rater reliability is poor. That means that if you were to film a person (for example) and then show it to lots of different clinicians, then those clinicians would disagree a lot as to

1) Whether the person had a mental illness
2) Which mental illness that person had

That makes it hard to see how heritable mental illness (or particular kinds of illness) are. One can only do linkage analysis with the diagnoses that people have been given...

It might be that the rates of mental illness have remained constant (or even that they have actually decreased) but now more people are presenting for help where they would have remained silent before.

The concept of mental illness has changed, however. It used to only be applied to people who needed to be confined (in asylums or by relatives) in order to stop them harming themselves or others. Nowdays it is certainly applied to people who are much much much less severe than that (though this isn't to undermine their suffering at all). It is just to say that once where we saw SPIRITUAL CRISIS or UNDERSTANDABLE RESPONSE TO UNFORTUNATE EVENTS we now see DEPRESSION and PROZAC and so on. 'Neurotic' cases (rather than 'psychotic') were originally treated by neurologists rather than psychiatrists. Middle class people who suffered from 'nerve complaints' were not thought to be suffering from mental illnesses. Whereas now, they are regarded as being psychiatric patients rather than neurological patitents. Or rather than as having a spiritual crisis.

The medicalization of (legitimate) distress / problems in living...

I see how it lines pockets. I don't really see how it helps...